Association of Ticagrelor vs Clopidogrel With Major Adverse Coronary Events in Patients With Acute Coronary Syndrome Undergoing
Percutaneous Coronary Intervention
Ricky D. Turgeon, BSc(Pharm), PharmD; Sheri L. Koshman, BSc(Pharm), PharmD; Erik Youngson, MMath;
Bryan Har, MD; Stephen B. Wilton, MD, MSc; Matthew T. James, MD, PhD; Michelle M. Graham, MD
IMPORTANCEGuidelines currently recommend ticagrelor over clopidogrel for patients with acute coronary syndrome (ACS) based on randomized clinical trial data in which ticagrelor reduced major adverse coronary events (MACE) vs clopidogrel but increased bleeding and dyspnea.
OBJECTIVETo compare the risk of MACE with ticagrelor vs clopidogrel in patients with ACS treated with percutaneous coronary intervention (PCI), to compare major bleeding and dyspnea, and to evaluate the association between P2Y12inhibitor adherence and MACE.
DESIGN, SETTING, AND PARTICIPANTSPopulation-based cohort study using data of patients discharged alive after PCI for ACS from the Alberta Provincial Project for Outcome Assessment in Coronary Heart Disease registry from April 1, 2012, to March 31, 2016, with follow-up to 1 year. Analysis began April 2018.
EXPOSURESOutpatient prescription for ticagrelor or clopidogrel within 31 days after PCI.
Adherence was defined as a medication refill adherence value of 80% or higher.
MAIN OUTCOMES AND MEASURESMajor adverse coronary events, a composite of all-cause death, hospitalization for ACS, unplanned coronary revascularization, or stent thrombosis within 365 days after index PCI. Secondary outcomes included hospitalization for major bleeding and emergency department visit for dyspnea.
RESULTSOf 11 185 individuals who underwent PCI, the median (interquartile range) age was 61 (54-71) years, and 2760 (24.7%) were women. Ticagrelor users (4076 [36.4%]) were generally younger and had fewer cardiac and noncardiac comorbidities than clopidogrel users. Ticagrelor was not associated with lower risk of MACE (adjusted hazard ratio [aHR], 0.97; 95% CI, 0.85-1.10); however, it was associated with an increased risk of major bleeding (aHR, 1.51; 95% CI, 1.29-1.78) and dyspnea (aHR, 1.98; 95% CI, 1.47-2.65). A total of 3328 ticagrelor users (81.6%) were adherent during the study vs 5256 of clopidogrel users (73.9%) (P < .001; χ2= 86.4). In the full cohort, adherence was associated with a lower MACE risk (aHR, 0.79; 95% CI, 0.69-0.90 for adherence ofⱖ80% vs <80%). Differences in other secondary outcomes were not statistically significant. Sensitivity and subgroup analyses were consistent with primary analyses.
CONCLUSIONS AND RELEVANCEIn this population-based cohort study of patients with ACS who underwent PCI, outpatient use of ticagrelor was not associated with a statistically significant reduction in MACE vs clopidogrel; however, it was associated with more major bleeding and dyspnea.
JAMA Intern Med. 2020;180(3):420-428. doi:10.1001/jamainternmed.2019.6447 Published online January 13, 2020.
Supplemental content
Author Affiliations: Department of Pharmacy, Vancouver General Hospital, Vancouver, British Columbia, Canada (Turgeon); Faculty of Pharmaceutical Sciences, University of British Columbia, Vancouver, British Columbia, Canada (Turgeon); Division of Cardiology, Faculty of Medicine and Dentistry, University of Alberta, Edmonton, Alberta, Canada (Koshman, Graham);
Alberta SPOR Support Unit, University of Alberta, Edmonton, Alberta, Canada (Youngson); Libin Cardiovascular Institute, University of Calgary, Calgary, Alberta, Canada (Har, Wilton); Division of Nephrology, Department of Medicine, University of Calgary, Calgary, Alberta, Canada (James).
Corresponding Author: Michelle M.
Graham, MD, University of Alberta, 8440 112th St NW, Ste 2C2 WMC, Edmonton, AB T6G 2B7, Canada (mmg2@ualberta.ca).
JAMA Internal Medicine | Original Investigation
G
uidelines recommend ticagrelor over clopidogrel as part of dual antiplatelet therapy with acetylsalicylic acid (or aspirin) in the treatment of individuals with acute coronary syndrome (ACS), regardless of plans for inva- sive management.1,2This recommendation is primarily based on the Platelet Inhibition and Patient Outcomes (PLATO) trial,3 in which patients with ACS receiving ticagrelor had a 16% rela- tive risk reduction in major adverse coronary events (MACE) and 22% relative risk reduction in all-cause death compared with those treated with clopidogrel. This improvement in ef- ficacy was counterbalanced by a 19% relative risk increase in major bleeding not associated with coronary artery bypass grafting and an 84% relative risk increase in dyspnea, which was generally reported as mild and transient.3-6Patients seen in routine practice may differ in several ways from those enrolled in clinical trials, including having higher risks of MACE and bleeding, more comorbidities, and lower likelihood of being prescribed and adhering to evidence- based therapies.7One registry study reported that 32% of con- secutively enrolled patients with myocardial infarction were ineligible for contemporary antiplatelet trials including PLATO.8Observational studies can complement clinical trials by evaluating the associations of interventions in representa- tive populations that reflect these factors and can identify barriers to replicating the results observed in randomized clinical trials.
The primary study objective was to assess the compara- tive association of ticagrelor and clopidogrel with reduced MACE using data from a population-based registry including all patients undergoing percutaneous coronary intervention (PCI) for ACS in a geographic region with universal access to health care. We hypothesized that ticagrelor would be asso- ciated with a lower risk of MACE than clopidogrel. Secondary objectives included the evaluation of the safety of ticagrelor compared with clopidogrel, as well as associations between adherence, persistence, and switching with MACE.
Methods
Data Sources
The Alberta Provincial Project for Outcome Assessment in Coronary Heart Disease registry prospectively collects detailed clinical data on all patients undergoing coronary angiography in Alberta, Canada (population of approxi- mately 4.3 million).9There are 3 sites where coronary angi- ography is performed that serve the entire province. The Alberta Provincial Project for Outcome Assessment in Coro- nary Heart Disease registry was used to define the study cohort, determine baseline clinical and procedural charac- teristics, and obtain data on death and coronary procedures.
The Pharmaceutical Information Network contains data on all outpatient prescriptions filled at pharmacies in Alberta, including fill dates and quantities. The Pharmaceutical Infor- mation Network was used to collect data on exposure to P2Y12inhibitors and other prescription medications, as well as data on P2Y12inhibitor adherence, persistence, and switching. Data on baseline hemoglobin concentration,
creatinine concentration, and estimated glomerular filtra- tion rate were obtained from the Alberta Health Services Laboratory Services. The Discharge Abstract Database and National Ambulatory Care Reporting System, which contain data on emergency department visits and hospitalizations in Alberta (eg, admission and discharge dates and International Statistical Classification of Diseases and Related Health Prob- lems, Tenth Revision [ICD-10] diagnosis codes) were used to identify history of atrial fibrillation (ICD-10 code I48.x recorded within 2 years prior to index) and study outcomes.
The University of Alberta Research Ethics Office approved this study with a waiver of informed consent because the data were deidentified when provided to the investigators.
Cohort Definition
Included participants were older than 18 years, underwent PCI for ACS between April 1, 2012, and March 31, 2016, were dis- charged alive from index hospitalization, and filled a first pre- scription for clopidogrel or ticagrelor within 31 days after un- dergoing PCI. Both clopidogrel and ticagrelor were approved for ACS in Canada during the entire study, and ticagrelor was available on the Alberta acute care formulary from October 1, 2012. Prevalent P2Y12inhibitor users, defined as individuals with a prescription for clopidogrel, prasugrel, or ticagrelor within 120 days prior to the index PCI, were excluded, as the present study was restricted to new users of P2Y12inhibitor therapy.
Exposure
Exposure was defined by a prescription fill in the Pharma- ceutical Information Network for clopidogrel or ticagrelor.
Three additional measures of exposure were calculated:
adherence, persistence, and switch, assuming an intended P2Y12inhibitor duration of 12 months according to guideline recommendations.1,2Adherence was estimated using the medication refill adherence (MRA),10,11defined as: MRA (%) = (total days’ supply within an interval / total days in interval) × 100, where the days’ supply for clopidogrel equaled the number of tablets filled in that time interval, and the days’ supply for ticagrelor equaled half the total tab- lets filled (because it is taken twice daily). Medication refill adherence was censored at time of death, such that adher- ence for patients who died prior to the next fill could not
Key Points
QuestionWhat is the association between ticagrelor vs clopidogrel and major adverse coronary events, major bleeding, and dyspnea in patients with acute coronary syndrome treated with percutaneous coronary intervention?
FindingsIn this cohort study of 11 185 patients, ticagrelor was not associated with a statistically significantly lower risk of major adverse coronary events compared with clopidogrel. However, it was associated with statistically significantly more major bleeding and dyspnea.
MeaningTicagrelor was not associated with a lower risk of major adverse coronary events in patients with acute coronary syndrome who underwent percutaneous coronary intervention.
exceed 100%. Patients with MRA of 80% or higher during the study were classified as adherent, as defined in previous ACS studies.3,12Participants were classified as persistent during the 365 days after the index event if they had gaps between P2Y12inhibitor prescription fills of less than the days’ supply plus a 15-day grace period.11Switch was defined as 1 or more prescription fills for a P2Y12inhibitor different from the first fill within 365 days after the index event. For patients who switched P2Y12inhibitors during the study, all P2Y12inhibitor fill information was considered for calcula- tion of adherence and persistence (ie, patients were consid- ered adherent if they filled the second P2Y12inhibitor at appropriate intervals).
During the study, administration of acetylsalicylic acid, 81 mg, daily, for all patients without contraindication was the standard of care after ACS in Alberta and was incorporated in all standardized preprinted order sets for ACS management and post-ACS care. Acetylsalicylic acid, 81 mg, is not rou- tinely recorded in the Pharmaceutical Information Network be- cause it is available without a prescription in Alberta. For the purposes of this study, all patients were assumed to receive ace- tylsalicylic acid, 81 mg, daily, during follow-up.
Outcomes
The primary outcome was the first occurrence of MACE, defined as a composite of all-cause death, hospitalization with nonfatal ACS (ICD-10 codes I20.0, I21, or I22 as most responsible diagnosis), coronary revascularization excluding planned staged PCI procedures (which is a prospectively filled data field in the Alberta Provincial Project for Outcome Assessment in Coronary Heart Disease registry), or stent thrombosis within 365 days after the index hospitalization.
Secondary outcomes included a composite of all-cause death, hospitalization with ACS, or ischemic stroke (ICD-10 codes I63.0 to I63.9 and I64); hospitalization for major b l e e d i ng ( I C D -10 c o d e s l i s te d i n e A p p e n d i x i n t h e Supplementas first or second diagnosis code); and emer- gency department visit for dyspnea (ICD-10 code R06.0 as most responsible diagnosis). The ICD-10 codes for ACS and major bleeding have been validated.13-15
Statistical Analyses
Baseline characteristics were summarized for each group as proportions for categorical variables and as medians with in- terquartile ranges for continuous variables. Proportions and medians were compared using the χ2and Mann-Whitney test, respectively. Hazard ratios (HR) and 95% confidence inter- vals for ticagrelor vs clopidogrel were obtained using Cox pro- portional hazard models, censored at time of P2Y12inhibitor switch. Models for nonfatal outcomes were also censored at time of death. To account for baseline differences, fully ad- justed Cox models were constructed for each outcome. Base- line characteristics, including all variables listed in Table 1, with a P value less than .20 in the univariable model, were in- cluded in a multivariable model; factors with a P value less than .05 in multivariable analysis remained in the final model. Ad- ditionally, the model was further adjusted for adherence using MRA as a continuous variable. To assess the association be-
tween adherence, persistence, and switching and outcomes, comparisons of MACE between ticagrelor and clopidogrel were stratified based on adherence of 80% or more vs less than 80%, persistence vs nonpersistence, and switching vs nonswitch- ing. Furthermore, risk of MACE was compared between those with adherence of 80% or more and less than 80%, persistent vs nonpersistent, and switchers vs nonswitchers.
To ensure robustness of the primary analysis, several sen- sitivity analyses were performed. First, ticagrelor and clopi- dogrel were compared in a propensity score–matched cohort, in which logistic regression was used to estimate the probabil- ity of ticagrelor use based on baseline variables listed in Table 1, and patients treated with ticagrelor were matched with those who received clopidogrel using 1:1 nearest-neighbor match- ing. Outcomes were compared using χ2tests and HRs from pro- portional hazards models stratified by matched pairs to ac- count for correlation between each matched pair. Second, the primary analysis was repeated with adjustment for a cardiac- specific comorbidity index.16Subgroup analyses for MACE, major bleeding, and dyspnea were performed for the follow- ing: age (<65, 65-74, or ≥75 years), diabetes, estimated glo- merular filtration rate (≥60 vs <60 mL/min), by fiscal year, with exclusion of patients receiving oral anticoagulants or with atrial fibrillation at baseline, exclusion of patients who switched P2Y12inhibitors, and high-risk criteria using 2 definitions:
PLATO inclusion criteria (defined as either ST-segment eleva- tion myocardial infarction, excluding those treated with res- cue PCI following failed fibrinolysis, or non–ST-segment elevation myocardial infarction plus any of the following:
age ≥60 years, multivessel coronary artery disease on coro- nary angiography, prior myocardial infarction, coronary artery bypass grafting, cerebrovascular disease, peripheral vas- cular disease, diabetes or renal disease, or estimated glomer- ular filtration rate <60 mL/min), and the TIMI Risk Score for Secondary Prevention.17The threshold for statistical signifi- cance was set at 2-sided P values less than .05. All analyses were performed using SAS, version 9.4 (SAS Institute) and R, version 3.4.3 (R Project for Statistical Computing). Analysis began April 2018.
Results
Cohort Characteristics
From April 1, 2012, to March 31, 2016, a total of 13 897 pa- tients underwent PCI for ACS in Alberta, Canada. Of these, 11 185 patients (80.5%) filled at least 1 prescription for a P2Y12inhibi- tor within 31 days of PCI and met all study eligibility criteria (Figure). Overall, clopidogrel was the most frequently pre- scribed P2Y12inhibitor during the study (7109 of 11 185 [63.6%]);
however, ticagrelor use steadily increased and was used in 3112 of 5523 patients (56.3%) by the second half of the study (Table 1 and eFigure in theSupplement).
The median (interquartile range) age was 61 (54-71) years, 2760 (24.7%) were women, and 4953 (44.3%) pre- sented with ST-segment elevation myocardial infarction.
Patients filled their first outpatient prescription for a P2Y12
inhibitor a median (interquartile range) of 3 (1-4) days after
Table 1. Baseline Clinical and Angiographic Characteristics and P2Y12Inhibitor Use During Study Follow-up
Characteristic
No. (%)
P Value Clopidogrel Group
(n = 7109)
Ticagrelor Group (n = 4076) Age, y
Median (IQR) 62 (54-72) 60 (53-69) <.001
≥75 1434 (20.2) 554 (13.6) <.001
Women 1831 (25.8) 929 (22.8) <.001
Fiscal yeara
2013 2593 (93.6) 177 (6.4)
<.001
2014 2105 (72.8) 787 (27.2)
2015 1311 (48.3) 1406 (51.7)
2016 1100 (39.2) 1706 (60.8)
ACS type
STEMI 3170 (44.6) 1783 (43.7)
.007
NSTEMI 2829 (39.8) 1683 (41.3)
Unstable angina 1081 (15.2) 576 (14.1)
Unknown 29 (0.4) 34 (0.8)
Cardiovascular history and risk factors
Prior MI 858 (12.1) 287 (7.0) <.001
Prior coronary artery bypass grafting 294 (4.1) 172 (4.2) .83
Prior PCI 951 (13.4) 478 (11.7) .01
Cerebrovascular disease 326 (4.6) 120 (2.9) <.001
Peripheral vascular disease 696 (9.8) 135 (3.3) <.001
Heart failure 434 (6.1) 126 (3.1) <.001
Atrial fibrillation 662 (9.3) 147 (3.6) <.001
Diabetes mellitus 1807 (25.4) 968 (23.7) .05
Hyperlipidemia 4163 (58.6) 2091 (51.3) <.001
Hypertension 4857 (68.3) 2532 (62.1) <.001
Smoking status
Current 2301 (32.4) 1060 (26.0)
<.001
Former 1580 (22.2) 715 (17.5)
Never/not recorded 3228 (45.4) 2301 (56.5)
Other comorbidities
Chronic pulmonary disease 714 (10.0) 147 (3.6) <.001
Renal disease 295 (4.1) 108 (2.6) <.001
Dialysis dependence 62 (0.9) 14 (0.3) .001
Liver disease 51 (0.7) 15 (0.4) .02
Malignancy 239 (3.4) 95 (2.3) .002
Coronary anatomy
1 Vessel 2395 (33.7) 1325 (32.5)
.54
2 Vessels 1849 (26.0) 1107 (27.2)
2 Vessels including proximal LAD 551 (7.8) 300 (7.4)
3 Vessels 1159 (16.3) 641 (15.7)
3 Vessels including proximal LAD 908 (12.8) 555 (13.6)
Left main 228 (3.2) 136 (3.3)
Minimal CAD 19 (0.3) 12 (0.3)
Ejection fraction, %
>50 2638 (37.1) 1298 (31.8)
<.001
35-50 1220 (17.2) 482 (11.8)
20-34 198 (2.8) 67 (1.6)
<20 19 (0.3) 4 (0.1)
Missing 3034 (42.7) 2225 (54.6)
Meets PLATO trial eligibility criteria 5639 (79.3) 3223 (79.1) .75
TIMI Risk Score for Secondary Prevention
Median (IQR) 2 (1-3) 1 (1-2) <.001
≥3 2014 (28.3) 770 (18.9) <.001
(continued)
index PCI and 1 (0-3) day after hospital discharge. Table 1 lists baseline characteristics by P2Y12inhibitor group.
Patients receiving ticagrelor were statistically significantly
younger, less likely to be women, and had a lower preva- lence of prior myocardial infarction, other cardiovascular disease, cardiac risk factors, and comorbidities compared with clopidogrel users. Ticagrelor users were also more likely to receive a drug-eluting stent than clopidogrel users and less likely to receive a prescription for an oral anticoagu- lant within 31 days after PCI.
Major Adverse Coronary Events
In unadjusted analyses, outpatient use of ticagrelor was associated with a lower risk of MACE than clopidogrel (HR, 0.84; 95% CI, 0.74-0.95) (Table 2). Unadjusted death rates were also lower with ticagrelor but not hospitalization for ACS, coronary revascularization or the composite of death, ACS, or ischemic stroke. After multivariable adjustment including baseline characteristics listed in Table 1 and adher- ence, differences in MACE (adjusted HR [aHR], 0.97; 95% CI, 0.85-1.10) and other cardiovascular outcomes were no longer statistically significant between ticagrelor and clopidogrel (Table 2).
Safety
In unadjusted analyses, ticagrelor was not associated with a statistically significantly greater risk of major bleeding com- pared with clopidogrel (HR, 1.09; 95% CI, 0.93-1.27). How- ever, ticagrelor was associated with a higher risk of major bleeding compared with clopidogrel in the fully adjusted model (aHR, 1.51; 95% CI, 1.29-1.78) (Table 2). This difference was mainly driven by an increase in gastrointestinal hemor- rhage (aHR, 2.02; 95% CI, 1.52-2.68) and pulmonary hemor- rhage (aHR, 1.49; 95% CI, 1.15-1.93). Furthermore, ticagrelor was associated with a statistically significantly greater risk of an emergency department visit for dyspnea, which persisted in the fully adjusted model (aHR, 1.98; 95% CI, 1.47-2.65) (Table 2).
Table 1. Baseline Clinical and Angiographic Characteristics and P2Y12Inhibitor Use During Study Follow-up (continued)
Characteristic
No. (%)
P Value Clopidogrel Group
(n = 7109)
Ticagrelor Group (n = 4076)
Stent placement 6747 (94.9) 3827 (93.9) .02
Drug-eluting stent 3989 (56.1) 3127 (76.7) <.001
Laboratory values, median (IQR)
Hemoglobin, g/dL 13.7 (12.5-14.8) 14.1 (13.0-15.1) <.001
Serum creatinine, mg/dL 0.95 (0.81-1.11) 0.95 (0.81-1.09) .41
eGFR, mL/min 62 (61-85) 81 (65-93) <.001
Other medications ≤31 d after PCI
Oral anticoagulant 1060 (14.9) 134 (3.3) <.001
Proton pump inhibitor 2630 (37.0) 1270 (31.2) <.001
Study P2Y12inhibitor utilization
MRA, median (IQR), % 98 (78-101) 99 (88-102) <.001
At month 0 to 6 100 (89-104) 101 (96-107) <.001
At month 7 to 12 95 (67-100) 95 (81-100) <.001
MRA ≥80% 5256 (73.9) 3328 (81.6) <.001
At months 0 to 6 5732 (80.6) 3602 (88.4) <.001
At months 7 to 12 4892 (68.8) 3048 (74.8) <.001
Persistence 4186 (58.9) 2579 (63.3) <.001
Switch 162 (2.3) 571 (14.0) <.001
Abbreviations: ACS, acute coronary syndrome; CAD, coronary artery disease; eGFR, estimated glomerular filtration rate; IQR, interquartile range; LAD, left anterior descending artery; MI, myocardial infarction;
MRA, medication refill adherence;
NSTEMI, non–ST-segment elevation myocardial infarction;
PCI, percutaneous coronary intervention; PLATO, Platelet Inhibition and Patient Outcomes;
STEMI, ST-segment elevation myocardial infarction;
TIMI, thrombolysis in myocardial infarction.
SI conversion factors: To convert hemoglobin to g/L, multiply by 10.0;
serum creatinine to μmol/L, multiply by 88.4.
aPercentages represent proportion of cohort initially prescribed that P2Y12inhibitor during the indicated fiscal year.
Figure. Cohort Derivation
13 897 Patients undergoing PCI for ACS between April 2012 and March 2016
11 854 P2Y12 inhibitor filled ≤31 d of PCI
11 185 Included in the cohort
2043 Excluded for not filling a P2Y12 inhibitor ≤31 d after index PCI 1541 Alive and discharged
302 Died before first fill ≤31 d 200 Not discharged within ≤31 d
669 Excluded
486 Filled P2Y12 inhibitor in 120 d prior to PCI 138 Coronary anatomy missing/
not determined/normal 45 First P2Y12 inhibitor filled
was prasugrel
7109 Received clopidogrel 5256 With P2Y12 inhibitor
adherence ≥80% (73.9%) 4186 With P2Y12 inhibitor
persistence to 1 y (58.9%) 159 Died within 1 y (2.2%)
4076 Received ticagrelor 3328 With P2Y12 inhibitor
adherence ≥80% (81.6%) 2579 With P2Y12 inhibitor
persistence to 1 y (63.3%) 63 Died within 1 y (1.5%)
Derivation of the study cohort from APPROACH between April 1, 2012, and March 31, 2016, is shown. Included patients from Alberta, Canada, who were 18 years or older who underwent percutaneous coronary intervention (PCI) for an acute coronary syndrome (ACS), were not receiving a P2Y12inhibitor in the 120 days preceding the ACS, and filled a prescription for clopidogrel or ticagrelor within 31 days after their PCI.
Association of Adherence, Persistence, and Switching With Major Adverse Coronary Events
Ticagrelor users were more likely to be adherent to P2Y12in- hibitor therapy during the entire study than clopidogrel users (3328 [81.6%] vs 5256 [73.9%]; P < .001; χ2= 86.4). Median MRA declined in both groups between months 0 to 6 and months 7 to 12, and 2579 ticagrelor users (63.3%) vs 4186 clopi- dogrel users (58.9%) persisted on a P2Y12inhibitor for 1 year (P < .001; χ2= 20.9) (Table 1). Major adverse coronary events
were not statistically significantly different between ticagre- lor and clopidogrel users regardless of adherence, persis- tence, and switching (Table 3). In the full cohort, MACE was lower in patients with P2Y12inhibitor adherence of 80% or higher compared with adherence under 80% (aHR, 0.79;
95% CI, 0.69-0.90).
Switching P2Y12inhibitors occurred in 571 ticagrelor us- ers (14.0%) vs 162 clopidogrel users (2.3%) (Table 1) and was not associated with an increased risk of MACE (aHR, 0.56;
Table 2. Association of Clopidogrel vs Ticagrelor With Outcomes Within 1 Year After Percutaneous Coronary Intervention for Acute Coronary Syndrome
Outcome
No. (%)
P Value
HR (95% CI) Clopidogrel Group
(n = 7109)
Ticagrelor Group
(n = 4076) Unadjusted
Adjusted for
Age and Sex Fully Adjusteda
MACE 828 (11.6) 419 (10.3) .03 0.84 (0.74-0.95)b 0.88 (0.78-0.995)b 0.97 (0.85-1.10)
All-cause death 159 (2.2) 63 (1.5) .01 0.73 (0.54-0.98)b 0.91 (0.67-1.23) 1.19 (0.87-1.62)
ACS 505 (7.1) 259 (6.4) .13 0.85 (0.73-1.00) 0.88 (0.75-1.03) 0.93 (0.79-1.09)
Coronary revascularization 338 (4.8) 178 (4.4) .35 0.83 (0.69-1.01) 0.85 (0.70-1.04) 0.90 (0.73-1.10)
PCI 249 (3.5) 133 (3.3) .50 0.87 (0.70-1.09) 0.88 (0.71-1.11) 0.87 (0.69-1.09)
Coronary artery bypass grafting 94 (1.3) 46 (1.1) .38 0.72 (0.49-1.06) 0.75 (0.51-1.11) 1.01 (0.68-1.52)
Stent thrombosis 25 (0.4) 19 (0.5) .35 1.30 (0.71-2.38) 1.29 (0.70-2.36) 1.66 (0.88-3.12)
Composite of death, ACS, or ischemic stroke
676 (9.5) 328 (8.0) .009 0.82 (0.72-0.94)b 0.87 (0.76-1.00) 0.93 (0.80-1.09)
Ischemic stroke 40 (0.6) 20 (0.5) .62 0.91 (0.52-1.57) 1.02 (0.59-1.78) 1.35 (0.75-2.40)
Major bleeding 449 (6.3) 277 (6.8) .32 1.09 (0.93-1.27) 1.23 (1.05-1.43)b 1.51 (1.29-1.78)b
Intracranial 11 (0.2) 5 (0.1) .67 0.68 (0.22-2.15) 0.71 (0.22-2.24) 0.88 (0.26-2.90)
Gastrointestinal 135 (1.9) 99 (2.4) .06 1.33 (1.02-1.74)b 1.55 (1.19-2.04)b 2.02 (1.52-2.68)b
Pulmonary 175 (2.5) 111 (2.7) .40 1.12 (0.88-1.43) 1.24 (0.97-1.58) 1.49 (1.15-1.93)b
Urologic 74 (1.0) 38 (0.9) .58 0.86 (0.58-1.30) 1.01 (0.67-1.52) 1.16 (0.76-1.78)
Other 85 (1.2) 42 (1.0) .43 0.89 (0.61-1.30) 0.99 (0.68-1.44) 1.18 (0.79-1.75)
Dyspnea 116 (1.6) 119 (2.9) <.001 1.80 (1.39-2.34)b 2.01 (1.54-2.61)b 1.98 (1.47-2.65)b
Abbreviations: ACS, acute coronary syndrome; HR, hazard ratio; MACE, major adverse coronary event; PCI, percutaneous coronary intervention.
aAdjusted for age, sex, and medication refill adherence as a continuous variable over the entire year and statistically significant variables from stepwise variable selection: MACE (prior PCI, diabetes, hyperlipidemia, coronary anatomy, drug-eluting stent, hemoglobin, creatinine, proton pump inhibitor
use); major bleeding (malignancy, hemoglobin, creatinine, oral anticoagulant use); dyspnea (fiscal year, prior myocardial infarction, heart failure, diabetes, chronic pulmonary disease, renal disease); and ischemic stroke (diabetes, hypertension, oral anticoagulant use).
bStatistically significant results.
Table 3. Major Adverse Coronary Events Within 1 Year After Percutaneous Coronary Intervention for Acute Coronary Syndrome Based on Adherence, Persistence, and Switching of Study P2Y12Inhibitor
Characteristic
MACE, No./Total No. (%) HR (95% CI)a
Full Cohort (N = 11 185)
Clopidogrel Group (n = 7109)
Ticagrelor Group (n = 4076)
Ticagrelor vs Clopidogrel Adherence
MRA ≥80% 881/8584 (10.3) 550/5256 (10.5) 311/3328 (9.9) 1.00 (0.86-1.16)
MRA <80% 366/2601 (14.1) 278/1853 (15.0) 88/748 (11.8) 0.88 (0.68-1.14) MRA ≥80% vs <80%,
HR (95% CI)a
0.79 (0.69-0.90)b 0.77 (0.66-0.89)b 0.86 (0.67-1.11) NA
Persistence
Yes 697/6765 (10.3) 433/4186 (10.3) 264/2579 (10.2) 1.02 (0.87-1.20)
No 550/4420 (12.4) 395/2923 (13.5) 155/1497 (10.4) 0.89 (0.73-1.09)
Persistent vs nonpersistent, HR (95% CI)a
0.90 (0.80-1.01) 0.85 (0.74-0.98)b 1.01 (0.81-1.24) NA
Switchc
Yes NA 6/92 (6.5) 47/521 (9.0) 1.49 (0.62-3.61)
No NA 752/6947 (10.8) 322/3505 (9.2) 0.98 (0.86-1.12)
Switch vs no switch, HR (95% CI)a
NA 0.56 (0.25-1.25) 0.88 (0.65-1.20) NA
Abbreviations: HR, hazard ratio;
MACE, major adverse coronary event;
MRA, medication refill adherence;
NA, not applicable.
aAdjusted for age, sex, prior percutaneous coronary intervention, diabetes,
hyperlipidemia, coronary anatomy, drug-eluting stent, hemoglobin, creatinine, and proton pump inhibitor use.
bStatistically significant results.
cExcludes switches occurring after MACE event.
95% CI, 0.25-1.25 with initial clopidogrel use and aHR, 0.88;
95% CI, 0.65-1.20 with initial ticagrelor use) (Table 3). Within 30 days after an emergency department visit for dyspnea, 28 of 112 ticagrelor users (25.0%) switched P2Y12inhibitors com- pared with 1 of 115 clopidogrel users (0.9%) (P < .001; χ2= 29.6).
Similarly, more ticagrelor users than clopidogrel users switched their P2Y12inhibitor within 30 days after a major bleed (18 of 255 [7.1%] vs 2 of 439 [0.5%]; P < .001; χ2= 25.1). Moreover, among ticagrelor users, switching occurred more frequently in those with an emergency department visit for dyspnea (54 of 119 [45.4%]) vs those without dyspnea (517 of 3957 [13.1%]) (P < .001; χ2= 100.1) and among those with hospitalization for major bleed (53 of 277 [19.1%]) vs those without major bleed (518 of 3799 [13.6%]) (P = .01; χ2= 6.5).
Sensitivity and Subgroup Analyses
Propensity score matching created a well-balanced cohort (n = 7422) with standardized differences for all baseline char- acteristics less than 0.1 (eTable 1 in theSupplement). Find- ings from the propensity score–matched analysis (Table 4) were consistent with the primary, multivariable-adjusted analy- ses, not demonstrating a statistically significant difference between ticagrelor and clopidogrel for MACE (HR, 1.00; 95%
CI, 0.86-1.17) but higher risks of major bleeding (HR, 1.52; 95%
CI, 1.24-1.87) and dyspnea (HR, 2.42; 95% CI, 1.70-3.45) with ticagrelor. Similarly, differences between ticagrelor and clopi- dogrel were not observed for MACE or any of its components when adjusted for a cardiac-specific comorbidity index (eTable 2 in theSupplement).
Subgroup analyses for MACE, major bleeding, and dysp- nea based on age, diabetes, estimated glomerular filtration rate, high-risk criteria based on PLATO or TIMI Risk Score for Sec- ondary Prevention criteria, fiscal year, and exclusion of pa- tients with atrial fibrillation, receiving oral anticoagulants, or switching P2Y12inhibitors were all consistent with results for
the overall cohort (eTable 3 in theSupplement). eTable 4 in the Supplementprovides the full details of the multivariate mod- els for MACE, major bleeding, and dyspnea. eTable 5 in the Supplementillustrates the change in the hazard ratio with step- wise addition of variables to the multivariate model for MACE.
Discussion
In this large, contemporary, population-based cohort study of patients who underwent PCI for ACS, outpatient use of ticagre- lor was not associated with a lower risk of MACE compared with clopidogrel; however, it was associated with a higher risk of ma- jor bleeding and dyspnea. Conversely, adherence to any P2Y12
inhibitor therapy was associated with 21% lower relative risk of MACE compared with P2Y12inhibitor nonadherence.
Our findings differ from prior studies on this topic, and this may be due to differences in methodology, patient popula- tions, and advances in interventional cardiology. The random- ized clinical trial that established the use of ticagrelor in ACS, PLATO,3was a multinational trial including all ACS subtypes regardless of planned invasive management, in whom PCI was performed predominantly using bare-metal stents and first- generation drug-eluting stents. Furthermore, outcome ascer- tainment in the PLATO trial included in-hospital outcomes. The large observational SWEDEHEART registry study,18which was also discordant with our findings, was not restricted to pa- tients with ACS who underwent PCI, did not specify the pro- portion of patients receiving second-generation drug-eluting stents, and defined exposure based on intended choice and du- ration of P2Y12inhibitor therapy on discharge without confir- mation of use and adherence using prescription fills or pa- tient interview. Like our study, the SWEDEHEART study18did not include in-hospital events and only included patients who survived to discharge. Conversely, our observational study was Table 4. Association of Clopidogrel vs Ticagrelor With Outcomes Within 1 Year After Percutaneous Coronary
Intervention for Acute Coronary Syndrome in Propensity Score–Matched Cohort
Outcome
No. (%)
P Value HR (95% CI) Clopidogrel Group
(n = 3711)
Ticagrelor Group (n = 3711)
MACE 368 (9.9) 380 (10.2) .64 1.00 (0.86-1.17)
All-cause death 54 (1.5) 61 (1.6) .51 1.10 (0.75-1.61)
ACS 228 (6.1) 235 (6.3) .74 1.02 (0.84-1.24)
Coronary revascularization 168 (4.5) 157 (4.2) .53 0.86 (0.67-1.09)
PCI 121 (3.3) 114 (3.1) .64 0.90 (0.68-1.19)
CABG 50 (1.3) 44 (1.2) .53 0.74 (0.47-1.15)
Stent thrombosis 7 (0.2) 18 (0.5) .03 2.57 (1.07-6.16)a
Composite of all-cause death, ACS, or stroke 290 (7.8) 299 (8.1) .70 1.02 (0.86-1.21)
Ischemic stroke 18 (0.5) 17 (0.5) .87 0.94 (0.48-1.86)
Major bleed 182 (4.9) 261 (7.0) <.001 1.52 (1.24-1.87)a
Intracranial 3 (0.1) 3 (0.1) >.99 1.00 (0.14-7.10)
Gastrointestinal 53 (1.4) 95 (2.6) <.001 2.10 (1.44-3.06)a
Pulmonary 81 (2.2) 105 (2.8) .08 1.32 (0.97-1.80)
Urologic 29 (0.8) 37 (1.0) .32 1.32 (0.79-2.22)
Other 32 (0.9) 38 (1.0) .47 1.29 (0.78-2.11)
Dyspnea 46 (1.2) 116 (3.1) <.001 2.42 (1.70-3.45)a
Abbreviations: ACS, acute coronary syndrome; CABG, coronary artery bypass grafting; HR, hazard ratio;
MACE, major adverse coronary event;
PCI, percutaneous coronary intervention.
aStatistically significant results.
restricted to patients with ACS who all underwent PCI, sur- vived to discharge, were new users of P2Y12inhibitor therapy, and primarily received second-generation drug-eluting stents, which have an improved safety profile, including a lower risk of stent thrombosis than bare-metal and older-generation drug- eluting stents.19The Dutch CHANGE DAPT cohort study, which enrolled patients with ACS who underwent PCI exclusively with second-generation drug-eluting stents, also found no statis- tically significant difference in MACE between ticagrelor and clopidogrel.20In aggregate, these study findings suggest that the increased potency of ticagrelor may not translate to im- proved efficacy in the era of second-generation drug-eluting stents, particularly after patients are hospital discharged.
A 2019 randomized clinical trial of patients with ACS under- going PCI found ticagrelor to be inferior to the once-daily potent P2Y12inhibitor prasugrel.21
Although the greater antiplatelet potency of ticagrelor did not translate to reduction in MACE in this study, it was asso- ciated with an increased risk of major bleeding. Given that pa- tients who underwent coronary artery bypass grafting during the index hospitalization were excluded from this study, these findings emulate the increased risk of non–coronary artery by- pass grafting–related major bleeding seen in PLATO.3More- over, our cohort included patients at higher risk of bleeding than those generally included in clinical trials, as indicated by a rate of major bleeding in the ticagrelor group that was higher than that reported in PLATO (6.8% vs 4.5%) despite the stricter bleeding definition used in our study and exclusion of events (including periprocedural bleeding) occurring during the in- dex hospitalization.3Similarly, patients in the ticagrelor group of this cohort had a 2-fold higher risk of emergency depart- ment visits for dyspnea vs those in the clopidogrel group.
Although ticagrelor-related dyspnea is generally mild and tran- sient, it persists and impairs quality of life in a subset of pa- tients, leading to increased health care utilization, nonadher- ence, and premature discontinuation.3-6,22
This study found that adherence to a P2Y12inhibitor was more strongly associated with risk of MACE than choice of the P2Y12inhibitor itself. These findings extend prior observa- tions that premature discontinuation of P2Y12inhibitors is as- sociated with a greater risk of death, rehospitalization, and stent thrombosis.23,24Several factors associated with lower adher- ence to P2Y12inhibitors,24including greater comorbidity bur- den and use of oral anticoagulation, were more prevalent in clopidogrel users within the present study. This may have ac- counted in part for the lower adherence and persistence in clopidogrel users compared with ticagrelor users. These re- sults should encourage clinicians to routinely ask patients whether they are taking their medications as prescribed and
identify and resolve barriers to adherence, including cost, ad- verse events (including dyspnea with ticagrelor), and burden from number or frequency of medications administered. The steady decline in adherence during the course of the present study, consistent with prior studies,25warrants ongoing as- sessment of medication adherence starting at hospital dis- charge and continuing at every follow-up visit.
Limitations
This study has limitations inherent to its observational design. First, residual unmeasured confounding may persist despite measurement and adjustment for a variety of known clinical, angiographic, and laboratory variables. Second, we excluded outcomes occurring during the index hospitaliza- tion, as information on in-hospital P2Y12inhibitor use was not available within our databases. Third, the definitions for exposure, adherence, and persistence assume that patients took their P2Y12inhibitors as filled and may overestimate true adherence. However, these definitions have been vali- dated, are consistent with those used in multiple prior stud- ies, and provide the closest surrogate to medication use available using administrative data.10,11Fourth, the adher- ence definition assumed that all patients were intended to receive a P2Y12inhibitor for at least 12 months, which was routine standard of care during the study. Fifth, we used all- cause death for the composite primary outcome because cause of death was not reliably ascertained within available databases. Therefore, these results are not directly compa- rable with the primary outcome of PLATO. However, we included the secondary outcome of all-cause death, ACS, or stroke, which was also evaluated in PLATO. Sixth, we used a novel outcome definition for dyspnea using a nonspecific ICD-10 symptom code that requires further validation. How- ever, post hoc analyses further evaluating this outcome revealed higher switch rates in the ticagrelor group among those with dyspnea and more switches from ticagrelor than clopidogrel within 30 days after an emergency department visit for dyspnea, supporting an association between ticagre- lor use and these events.
Conclusions
In a large, representative population-based cohort of patients who underwent PCI for ACS primarily using second- generation drug-eluting stents, ticagrelor was not associated with a lower risk of MACE compared with clopidogrel; how- ever, it was associated with a higher risk of major bleeding and emergency department visits for dyspnea.
ARTICLE INFORMATION
Accepted for Publication: November 2, 2019.
Published Online: January 13, 2020.
doi:10.1001/jamainternmed.2019.6447
Author Contributions: Dr Youngson had full access to all of the data in the study and takes
responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Turgeon, Koshman, Har, Wilton, Graham.
Acquisition, analysis, or interpretation of data:
Turgeon, Koshman, Youngson, Wilton, James, Graham.
Drafting of the manuscript: Turgeon, Youngson, Har.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Turgeon, Youngson.
Obtained funding: James
Administrative, technical, or material support:
Turgeon, Wilton, James, Graham.
Supervision: Koshman, Har, Graham.
Conflict of Interest Disclosures: Dr Har reports personal fees from Bayer, AstraZeneca, Novartis, and Abbott Laboratories during the conduct of the study. Dr Wilton reports grants from Medtronic Canada, Boston Scientific, and Abbott Vascular and
personal fees from Arca Biopharma outside the submitted work. Dr James reports grants from Amgen Canada outside the submitted work.
No other disclosures were reported.
REFERENCES
1. Mehta SR, Bainey KR, Cantor WJ, et al; members of the Secondary Panel. 2018 Canadian Cardiovascular Society/Canadian Association of Interventional Cardiology focused update of the guidelines for the use of antiplatelet therapy. Can J Cardiol. 2018;34(3):214-233. doi:10.1016/j.cjca.
2017.12.012
2. Levine GN, Bates ER, Bittl JA, et al. 2016 ACC/AHA guideline focused update on duration of dual antiplatelet therapy in patients with coronary artery disease: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines: an update of the 2011 ACCF/AHA/SCAI guideline for percutaneous coronary intervention, 2011 ACCF/AHA guideline for coronary artery bypass graft surgery, 2012 ACC/AHA/ACP/AATS/PCNA/SCAI/STS guideline for the diagnosis and management of patients with stable ischemic heart disease, 2013 ACCF/AHA guideline for the management of ST-elevation myocardial infarction, 2014 AHA/ACC guideline for the management of patients with non–ST-elevation acute coronary syndromes, and 2014 ACC/AHA guideline on perioperative cardiovascular evaluation and management of patients undergoing noncardiac surgery. Circulation. 2016;
134(10):e123-e155. doi:10.1161/CIR.
0000000000000404
3. Wallentin L, Becker RC, Budaj A, et al; PLATO Investigators. Ticagrelor versus clopidogrel in patients with acute coronary syndromes. N Engl J Med. 2009;361(11):1045-1057. doi:10.1056/
NEJMoa0904327
4. Storey RF, Becker RC, Harrington RA, et al.
Characterization of dyspnoea in PLATO study patients treated with ticagrelor or clopidogrel and its association with clinical outcomes. Eur Heart J.
2011;32(23):2945-2953. doi:10.1093/eurheartj/ehr231 5. Storey RF, Bliden KP, Patil SB, et al;
ONSET/OFFSET Investigators. Incidence of dyspnea and assessment of cardiac and pulmonary function in patients with stable coronary artery disease receiving ticagrelor, clopidogrel, or placebo in the ONSET/OFFSET study. J Am Coll Cardiol. 2010;56 (3):185-193. doi:10.1016/j.jacc.2010.01.062 6. Bonaca MP, Bhatt DL, Cohen M, et al;
PEGASUS-TIMI 54 Steering Committee and
Investigators. Long-term use of ticagrelor in patients with prior myocardial infarction. N Engl J Med. 2015;372(19):1791-1800. doi:10.1056/
NEJMoa1500857
7. Kennedy-Martin T, Curtis S, Faries D, Robinson S, Johnston J. A literature review on the
representativeness of randomized controlled trial samples and implications for the external validity of trial results. Trials. 2015;16(1):495. doi:10.1186/
s13063-015-1023-4
8. Udell JA, Wang TY, Li S, et al. Clinical trial participation after myocardial infarction in a national cardiovascular data registry. JAMA. 2014;
312(8):841-843. doi:10.1001/jama.2014.6217 9. Ghali WA, Knudtson ML. Overview of the Alberta provincial project for outcome assessment in coronary heart disease: on behalf of the APPROACH investigators.Can J Cardiol. 2000;16 (10):1225-1230.
10. Hess LM, Raebel MA, Conner DA, Malone DC.
Measurement of adherence in pharmacy administrative databases: a proposal for standard definitions and preferred measures. Ann Pharmacother. 2006;40(7-8):1280-1288.
doi:10.1345/aph.1H018
11. Andrade SE, Kahler KH, Frech F, Chan KA.
Methods for evaluation of medication adherence and persistence using automated databases.
Pharmacoepidemiol Drug Saf. 2006;15(8):565-574.
doi:10.1002/pds.1230
12. Rasmussen JN, Chong A, Alter DA. Relationship between adherence to evidence-based
pharmacotherapy and long-term mortality after acute myocardial infarction. JAMA. 2007;297(2):
177-186. doi:10.1001/jama.297.2.177
13. McCormick N, Lacaille D, Bhole V, Avina-Zubieta JA. Validity of myocardial infarction diagnoses in administrative databases: a systematic review. PLoS One. 2014;9(3):e92286. doi:10.1371/journal.pone.
0092286
14. Gomes T, Mamdani MM, Holbrook AM, Paterson JM, Hellings C, Juurlink DN. Rates of hemorrhage during warfarin therapy for atrial fibrillation. CMAJ. 2013;185(2):E121-E127.
doi:10.1503/cmaj.121218
15. Arnason T, Wells PS, van Walraven C, Forster AJ.
Accuracy of coding for possible warfarin complications in hospital discharge abstracts.
Thromb Res. 2006;118(2):253-262. doi:10.1016/
j.thromres.2005.06.015
16. Azzalini L, Chabot-Blanchet M, Southern DA, et al. A disease-specific comorbidity index for
predicting mortality in patients admitted to hospital with a cardiac condition. CMAJ. 2019;191(11):
E299-E307. doi:10.1503/cmaj.181186 17. Bohula EA, Bonaca MP, Braunwald E, et al.
Atherothrombotic risk stratification and the efficacy and safety of vorapaxar in patients with stable ischemic heart disease and previous myocardial infarction. Circulation. 2016;134(4):
304-313. doi:10.1161/CIRCULATIONAHA.115.019861 18. Sahlén A, Varenhorst C, Lagerqvist B, et al.
Outcomes in patients treated with ticagrelor or clopidogrel after acute myocardial infarction:
experiences from SWEDEHEART registry. Eur Heart J. 2016;37(44):3335-3342. doi:10.1093/eurheartj/
ehw284
19. Palmerini T, Biondi-Zoccai G, Della Riva D, et al.
Stent thrombosis with drug-eluting and bare-metal stents: evidence from a comprehensive network meta-analysis. Lancet. 2012;379(9824):1393-1402.
doi:10.1016/S0140-6736(12)60324-9 20. Zocca P, van der Heijden LC, Kok MM, et al.
Clopidogrel or ticagrelor in acute coronary syndrome patients treated with newer-generation drug-eluting stents: CHANGE DAPT. EuroIntervention.
2017;13(10):1168-1176. doi:10.4244/EIJ-D-17-00634 21. Schüpke S, Neumann F-J, Menichelli M, et al;
ISAR-REACT 5 Trial Investigators. Ticagrelor or prasugrel in patients with acute coronary syndromes. N Engl J Med. 2019;381(16):1524-1534.
doi:10.1056/NEJMoa1908973
22. Gaubert M, Laine M, Richard T, et al. Effect of ticagrelor-related dyspnea on compliance with therapy in acute coronary syndrome patients. Int J Cardiol. 2014;173(1):120-121. doi:10.1016/j.ijcard.
2014.02.028
23. Spertus JA, Kettelkamp R, Vance C, et al.
Prevalence, predictors, and outcomes of premature discontinuation of thienopyridine therapy after drug-eluting stent placement: results from the PREMIER registry. Circulation. 2006;113(24):
2803-2809. doi:10.1161/CIRCULATIONAHA.106.
618066
24. Airoldi F, Colombo A, Morici N, et al. Incidence and predictors of drug-eluting stent thrombosis during and after discontinuation of thienopyridine treatment. Circulation. 2007;116(7):745-754.
doi:10.1161/CIRCULATIONAHA.106.686048 25. Czarny MJ, Nathan AS, Yeh RW, Mauri L.
Adherence to dual antiplatelet therapy after coronary stenting: a systematic review. Clin Cardiol.
2014;37(8):505-513. doi:10.1002/clc.22289
